I have regressed real interest rates on themselves lagged 1 month ARIMA(1,1,0).  The box pierce q-stat is lower than the critical chi-square value (90% confidence level) for 38 degrees of freedom, and the sample autocorrelations go to zero after 1 lag.  It looks like this will be a pretty good model.  However, I am supposed to compare other models, like an ARIMA(2,1,0) or an ARIMA(0,1,1).  I don't know how to get the phi 1 and phi 2 for the ARIMA(2,1,0) model or the theta1 for the ARIMA(0,1,1) model. 

Neas, please give me some guidance here.

[NEAS: Take first differences to get an AR(2) or an MA(1) model. Use multiple linear regression for AR(2) and Yule Walker equations for MA(1).]